Modular Knee Brace

Knee support tends to be either too generic (one-size sleeves with undifferentiated compression) or too rigid (post-operative braces that trade usable range of motion for immobilisation). The design gap is a brace whose parts can be swapped or resized per user and per recovery stage, without replacing the whole device — so modularity is the core premise, not a feature.

It started from a specific user: the brace was sized for one person with long-term knee instability after ligament reconstruction, aimed at early-stage rehabilitation and gait support.

R2–R5 need only calipers, a goniometer, a stopwatch, and a scale once the parts are printed — the cheapest verification data in this portfolio.

Phase 1 was my individual design project for Design for Manufacturing (EGN 3433C, Spring 2025): design, CAD, FEA, and report. Phase 2 is independent work.

• Phase 1: FSR array at the brace–limb interface, signal conditioning, Arduino, and a Python pipeline from raw ADC counts through a calibration transform to gait-cycle plots. The physical prototype no longer exists.
• SolidWorks part set and assembly, intended for carbon-fibre-reinforced nylon. Linear static FEA in SolidWorks Simulation on three models: the hinge-and-connector assembly, the right upper connector, and the outer right hinge, plus a SimulationXpress study of the right lower connector. All were modelled as Nylon 101, because the software has no carbon-fibre-nylon material, under loads of 0.25–6 lb.
• Fusion 360 rebuild: parametric thigh and calf frames (thigh bore 5.75 in, calf bore 4.50 in, named parameters for fit, wrap angle, socket, and strap path), exact copies of the original connectors and hinges positioned on the knee axis, and Python Fusion-API scripts to build, measure, and export parts.
• Print-ready STL exports in millimetres.

• Fit checked in CAD at every joint of the 10-part assembly; corrections to the original parts are documented in the repo.
• FEA, right lower connector (SimulationXpress): Nylon 101 (yield 60 MPa), 3 lbf applied, fixed on 15 faces, standard solid mesh. Max von Mises stress 35.7 MPa, max displacement 2.96 mm, minimum factor of safety 1.68 — below the 2–4 target. The report’s conclusion: not yet a safe design for a wearer.
• FEA, hinge-and-connector assembly: the revised run peaked just under nylon’s ~11,500 psi yield at loads of a few pounds — effectively no margin.

PendingPhysical verification waits on the print: caliper deviation against CAD nominal, goniometer range of motion, swap time, and mass. The Phase 1 FSR calibration data is being located; until it is, no load or calibration figure is claimed.

• The Phase 1 hardware and sensing layer no longer exist. A prototype whose only record is the physical object is a prototype you will lose. Phase 2’s response — parametric CAD, version-controlled scripts, documented design intent — is the structural fix.
• The first assembly mesh failed. Clearances between the hinges and connectors were too tight, leaving interferences the mesher couldn’t resolve. I bonded the contacts to get a solution, which ran about 30 minutes — and bonding hides exactly the interfaces most likely to fail.
• A second assembly run reported a factor of safety around 5,000 under the same conditions. A number that large points to a loading or boundary-condition error, not an overbuilt part, so I don’t treat that run as a result.
• No study met the 2–4 safety-factor target. The design needs geometry changes at the connectors and hinges, and that is part of why Phase 2 rebuilds the model parametrically rather than patching it.
• The rebuild found inherited issues in the originals: about 0.003 in of press fit on one side of each hinge seat, and connectors and hinges that don’t follow the frame parameters because they are copied geometry.

• Not a medical device, not intended for patient use, and not evaluated under any FDA pathway. No human-subject testing.
• Phase 1 measured brace–limb interface contact pressure — a proxy for load transfer through the brace, not tibiofemoral joint force, which can’t be measured non-invasively.
• Static linear FEA only, with isotropic Nylon 101 standing in for the intended carbon-fibre nylon (and for the PLA/PETG of FDM prints). No fatigue, soft-tissue compliance, strap slip, or out-of-plane loading. The applied loads (0.25–6 lb) haven’t yet been justified against gait loading, mesh convergence hasn’t been checked, and the Phase 2 Fusion geometry hasn’t been re-simulated.
• Sized for one user’s leg (n = 1 geometry).
• Materials haven’t been evaluated for skin contact (ISO 10993), and durability is uncharacterised.